Enhanced Detection Techniques based on Distance Estimation with Radar for Human Bio-signals

Yun, Younguk Kwangwoon University Graduate School 2020 국내박사

This dissertation aims to increase the detection performance of distance estimation exploiting radar-based technology. We provide the various signal processing schemes to improve distance and detection performance through various signal processing of time-series signals and data collected by radar. There are two types of radar that we exploited, impulse radio ultra-wideband (IR-UWB) and Doppler radar. It has verified that are distance detection scheme and vital sign detection scheme. The former one is studied by exploiting IR-UWB radar and the latter one is studied by exploiting IR-UWB and Doppler radar. we limit the scope of vital signs that are heartbeat and respiration rate. Recently, it has been developed that are various technologies such as hyper-connected society, the 4th industrial revolution, cyber-physical system (CPS), and on-demand and expanded beyond the IoT and IoE epoch. Accordingly, it is emerging as a core technology that are location estimation, localization, or indoor positioning technology. In relation to the 4th Industrial Revolution, the indoor positioning system (IPS) is needed to complete the technologies used in industrial sites such as smart factories of CPS, smart city, smart building, and smart environment. In the case of outdoor, the location tracking technology is mainly based on the GPS system, but indoor and factory cannot exploit GPS and requires more accurate location tracking technology in the indoor environments, so various IPS are studied by using various sensors. However, it is being actively conducted researches on bio-signal detection schemes with increasing interest in healthcare along with positioning technology utilized in service aspects such as on-demand and fin-tech. We study the distance tracking technique that estimates the location or position exploiting the radar. And we discuss and proposed schemes to improve the performance of signals measured using radar through signal processing algorithms and verify the proposed scheme through experiments. The IR-UWB radar uses the 6.0-8.5GHz frequency band and the Doppler radar uses the 24GHz frequency band. For the distance estimation experiment, the experiment was conducted on the 6th floor of Kwangwoon University's Hwado-building using IR-UWB to move back and forth to the wide corridor 12m and the narrow corridor 15m. the movement speed was 0.5m/s. Bio-signal was measured using IR-UWB and Doppler radar. First, in the distance estimating experiment, the radar was installed at the height of the chest position of the subject and measured at a distance of 50cm from the subject. The experiment is repeated 30times for 1 minute with a total of 5 subjects. Bio-signal detection experiments using Doppler radar were conducted in general offices and vehicles. It is divided into the case with and without the subject's movement when the bio-signal experiment measure. The absence of the subject's movement includes both measurements with a distance of 0.5 to 2m and a case where only the vehicle is turned on. The motion test is conducted that is driving situation on the straight section of a normal highway. Various signal processing schemes are used that are singular value decomposition, low-rank approximation, and cross-correlation for estimating distance using IR-UWB. In the case of distance estimating scheme, the performance is increased by about 12 times compared with the conventional scheme. The conventional scheme has a maximum of 10m, an average error of 2.257m, and the proposed scheme has a maximum distance of 15m and an average error of 0.22m. In the respiratory detection study exploiting IR-UWB, the estimated performance is approximately 94.02% using Gram Schmidt orthogonalization techniques. In the case of using Doppler radar, the data are examined on the frequency axis through fast Fourier transform. The respiratory detection accuracy at stop status was 93.75%, heart rate 96.64%, the respiration detection accuracy at driving was 68.75%, and heart rate 84.48%. Various signal processing techniques can be utilized to increase the accuracy of driving.

Novel enzymatic synthesis of phenethyl formate using lipase

Shin, Min Guk Kwangwoon University Graduate School 2019 국내석사

최근 다양한 산업에서 발생하는 부생가스인 C1 가스를 유용화합물로 전환하는 기술이 세계적으로 주목을 받고 있다. 버려지는 C1 가스 중 일산화탄소를 유용화합물인 포름산으로 변환하는 기술이 특히 각광받고 있다. 그러나 포름산의 낮은 가격으로 인한 공정상의 경제성 확보를 위한 추가적인 고부가가치화가 필요하다. 따라서 포름산의 고부가가치화를 위한 물질중 하나로 포름산 에스터를 합성하는 다양한 방법이 적용되고 있다. 에스터는 산업에서 중요한 방향족 물질로서 화장품, 향수 등에 널리 사용된다. 현재는 포름산 에스터를 비롯한 다양한 에스터를 식물에서 추출하거나 화학적으로 합성하기 때문에 낮은 생산성과 환경오염이 문제되고 있다. 이를 극복하기 위한 방안으로 효소를 이용한 포름산 에스터 합성이 적용 가능하다. 다양한 에스터 합성에 효소가 적용되어 오고 있으나, 효소 기반의 포름산 에스터를 합성하는 연구는 진행된 바가 없다. 따라서 본 연구에서는 포름산 에스터 중 향료산업에서 널리 쓰이는 페네틸 포름산을 고정화 효소(lipase)를 통하여 합성하고 최적화하는 실험을 진행하였다. 효소선정, 효소 농도, 몰 비율, 온도, 용매에 대한 최적화를 진행하였으며 Novozym 435, 15g/L, 1:5, 40℃, 1,2-dichloroethane에서 최대 전환율인 95.92%를 확인하였다. 동일 조건에서 용매가 톨루엔인 경우에는 93.18%의 최대 전환율을 보였다. 또한 공정상의 경제성을 고려하여 효소를 20회 재사용 한 결과, 최소 92% 이상의 전환율을 보였다. 본 연구에서는 포름산과 페네틸 알코올의 직접적인 에스터화로 페네틸 포름산을 합성하였으며, 부산물로 오직 물만 생성되므로 친환경적이라 할 수 있다. 포름산을 50배에서 100배 수준으로 고부가가치화를 할 수 있고, 현재 상용화된 에스터를 대체할 수 있으므로 상업적 적용 가능성이 크다고 할 수 있다.

Study on graphene-based flexible and stretchable supercapacitors for wearable applications

Kim, Jiyoung Kwangwoon University Graduate School 2019 국내석사

Energy storage devices, which are suitable for wearable electronic devices including smartwatches have been attracting attention. Among various sorts of energy storage devices, flexible supercapacitors are considered to be promising devices for the next generation due to its long cycle life and chemical stability. Also, the flexible supercapacitors can be fabricated into various shapes easily depending on users' demands and practical applications. However, the durability, stretchability, physical and bio-compatible properties of the supercapacitors comprised of electrodes, electrolyte, and packaging materials have been challenged for their successful commercialization. In this research, to address the demands of issues as mentioned above, two kinds of Graphene-based flexible and stretchable supercapacitors have been designed, fabricated, and characterized for wearable electronic applications. Firstly, graphene flakes (GFs)-coated carbon fibers (CFs)-based flexible cable-shaped supercapacitor has been studied and developed. The graphene flakes on CFs surface are highly helpful to enlarge specific surface area and electrical conductivity of the electrodes, which affect higher specific capacitance. As-fabricated cable-shaped flexible supercapacitor has been analyzeed electrochemically by using cyclic voltammetry (CV) and galvanostatic charging-discharging (GCD) techniques. The fabricated supercapacitor exhibited the rectangular shaped CV curve, which shows the electric double-layer capacitor (EDLC) property. The specific capacitance of 15.099~6.492mF/cm2 and energy density of 2.097~0.902µWh/cm2 have been achieved in the scan rate range of 50~300mV/s, respectively. These values are about 1.9 times larger than those of the devices without being coated with GFs on the CF surface. In addition, the specific capacitance showed only 3.4% of a difference between the straight and twisted position, which shows the mechanical stability of the developed flexible supercapacitor. Secondly, a laser-induced graphene-based stretchable supercapacitor has been studied and developed. A simple and low cost transferring process of graphene into polydimethylsiloxane (PDMS) substrate were applied to improve the specific surface area and specific capacitance of the electrodes. Also, the applied transferring process was highly effective to obtain the flexible and stretchable mechanical feature of the supercapacitor. The deposited AgNWs formed a network structure below the LIG layer enhanced the conductivity of the electrode about 5.5 times and assured to maintain the electrical performance of the supercapacitor during physical deformation. NaCl and agarose were selected as an electrolyte because of its non-toxic properties. From the electrochemical analysis, the rectangular shape of the CV curves was also observed. The specific capacitance of 1.319~0.618 mF/cm2 and energy density of 0.183~0.086µWh/cm2 were calculated in the scan rate range of 50~500mV/s, respectively. From the stability test, the capacitance maintained about 95% of the initial value after 1000 cycles of charging and discharging repetition. From the mechanical deformation test, 80% of the capacitance in the initial state was observed at 20% of stretching elongation and 135° of bending angle. These results showed that the flexible and stretchable supercapacitors developed in this study are one of the promising candidates among various energy storage devices for wearable electronic applications.

Effect of task virtuality on turnover intention through workplace loneliness moderated by Self-construal

Suh, Gene Hyung Kwangwoon University Graduate School 2020 국내석사

The rapid advancement of technological and flexible work environment has made virtual work globally prevalent in all levels of organizations. As research on the dynamics of virtuality have been widely elaborated, there are still many unknowns about the impact of task virtuality at the individual level that needs to be understood, specifically related to turnover intention. This research investigates workplace loneliness as the link between task virtuality and turnover intention, dependent upon workers' orientation of their self-construal and its interaction effect on turnover intention. Integrating the affordance theory and the affect theory of social exchange, this study aims to shed light on the central role of workplace loneliness in explaining the task virtuality-turnover intention relationship, predicting that despite virtually connected employee's desire to connect with others, lack of opportunity for constructing a quality social and working relationship is associated with higher turnover intention due to the increased feelings of loneliness at work. The moderating effect of self-construal on workplace loneliness was also examined. Analysis of 300 office workers from a diverse field of employment shows that employees high on task virtuality and workplace loneliness are more likely to leave the organization. However, the mediating effect of loneliness and the moderating effect of self-construal was not found.

Influence of hole transport layers on the power output stability of perovskite solar cells

Hong, Jungyun Kwangwoon University Graduate School 2020 국내석사

Although many researches have been studying to make high-efficiency perovskite solar cells, various unstable characteristics have to be solved for commercialization of the device. In addition, the process method to reduce the process cost and the price of the used materials are important factors. When Nickel oxide (NiOx) is used as a hole transport layer, the output instability such as hysteresis and light-soaking effects was not shown by measuring the J-V characteristics. Furthermore, the incorporation of an ultra-small amount of polymer additives, poly(ethylene) tridecyl ether (PTE) which can retard the crystallization of perovskite by mixing with perovskite precursors reduced roughness, trap sites, and photogenerated charge carriers’ loss in the perovskite films. Atomic force microscopy (AFM), space charge limited current (SCLC) for hole only devices, light intensity dependent open-circuit voltage and short-circuit current measurements support these results. As a result, the fabricated device improved efficiency and operational stability. Finally, the microwave annealing method is adopted instead of the high temperature thermal annealing method which is commonly used to form NiOx films from the NiOx 고효율 페로브스카이트 태양 전지를 만들기 위한 많은 연구가 진행중에 있 지만 소자의 상용화를 위해서 여러가지 불안정한 특성들이 해결되어야 할 과제로 남겨져 있다 . 뿐만 아니라 공정단가를 줄이기 위한 공정방법 및 사용되는 물질들의 가격도 중요하게 생각해야할 부분이다 본 연구 에서는 p i n 구조의 페로브스카이트 태양전지에서 , 사용되어진 정공수송층에 따라 발생 완화되는 출력효율 불안정성에 대하여 관찰하였고 안정성이 향상된 소자 개발을 위한 실험과 연구가 진행되었다 . copper iodide CuI 를 정공수송층으로 사용한 경우 , 투명전극으로 사용한 Indium tin oxide (ITO) 표면의 물리적 결점을 따라 CuI 의 큰 입자 뭉침에 의해 발현된 아이오 딘 이온 (I 의 이동현상으로 인하여 출력효율 불안정성이 생성되었음을 여러 실험을 통하여 확인하였다 또한 소자 에이징 동안에 이루어진 아이오딘 이온 의 확산 이동현상을 통하여 페로브스카이트층과 전자수송층인 [ phenyl C 61 butyric acid methyl ester (PCBM) 계면에 축적된 아이오딘 이온이 불안정성을 크게 유발함을 의도적으로 제작된 소자를 제작함으로써 관찰하였다 여기서 진행된 실험 으로는 소자 에이징 시간 에 따른 , 전기적 전제조건에 따른 , 스캔 속도에 따른 J V 특성 측정 빛의 세기에 따른 개방전압 측정 이다 이 원인을 제어하기 위하여 CuI 아래층에 CuI 의 큰 입자 뭉침을 제어 하기위한 다른 정공수송층을 도입 함으로써 출력효율 불안정성이 해결되어진 소자를 개발하였다 Nickel oxide ( NiO x 를 정공 수송층 으로 사용했을 경우에는 위와 같은 불안정성을 보이지 않았음을 J V 특성 측정 을 통하여 확인하였고 , 추가 연구로 페로브스카이트의 결정화를 지연시키는 역할을 하는 고분자 첨가제 , p oly(ethylene glycol) tridecyl ether (PTE) 를 페로브스카이트 층 에 극소량 첨가함으로써 roughnessroughness와 trap site , photogenerated charge carriercarrier의 손실 을 줄였음을 atomic force microscopy ( 측정과 hole only device device의 space charge limited current (SCLC) 측정 , 빛의 세기에 따른 개방전압 및 단락전류 측정들을 통하여 확인할 수 있었다있었다. 이로 인하여 제작된 소자는 효율 향상과 작동안정성이 증가함을 보였다. 마지막으로는, NiONiOx 프리커서를 열처리 하기위해 보통 많이 사용되는 고온 열 어닐링 방법 대신 마이크로웨이브 어닐링 방법을 채택하였다채택하였다. X-ray photoelectron spectroscopy (XPS) 와 ultraviolet photoelectron spectroscopy (UPS) 를 통해 NiO x의 형성 유무와 HOMO level level을 알 수 있었으며 , J-V 특성 측정과 파장에 따른 external quantum efficiency (EQE) 측정 , NiO x의 저항을 측정함으로써 증가한 단락전류를 확인 하였다하였다. NiO x층의 마이크로웨이브 어닐링을 통하여 공정시간을 효과적으로 단축시키 며 효율 도 향상시킨 페로브스카이트 소자가 제작되었다.

Scalable Cooperative Relay Scheme for Visible Light Communication Systems

Park, Inhwan Kwangwoon University Graduate School 2012 국내석사

(A) Crosscultural study on the influence of salespersons' non-verbal communication on the purchase intentions of southeastern tourist to Korea

Masniati Kwangwoon University Graduate School 2021 국내석사

This study aims to determine the effect of non-verbal salesperson communication on consumer buying intentions. This study uses a research design developed through the scenario method. The model framework and components are designed based on the real conditions and ratings of tourists in Korea who have a high tendency to make purchases. This development procedure consists of: (1) observation research to develop scenarios, and (2) a survey method uses the scenarios developed on the effectiveness of non-verbal communication. For the scenario development, a series of systematic observation was executed at several cosmetic stores in Seoul, South Korea. Through the observations, the facial expressions, body language, responding attitude, and proactive behaviors of salespersons were extracted for the final scenario. To validate the hypotheses, respondents were allocated to either scenario A (76) and B (70). With the Hotelling's Trace MANOVA test, the results show that scenario A and B elicited significantly different responses for most of the dependent measurements, which are willing to visit again, willing to spend more money, willing to spend a long time shopping at the store, willing to recommend and even tent to trust the quality of the shop when the salesperson provides proactive non-verbal communication such as kind facial expression. listening sincerely for questions and even willing to explain beyond the questions about goods. Moreover, the results also show that cultural factors play an important role in nonverbal communication which is often applied but neglected in buying and selling transaction studies In this study, the influence on their purchase intention was significantly different across the nationality of the tourists in that Malaysia, Indonesia, Thailand and the Philippines have different cultural backgrounds such as religious orders, beliefs, or languages. The results of the study not only provide an understanding that salespersons’ role in various forms of communication messages and attitudes towards the customer but also each distinct cultural differences embedded among seemingly homogeneous southeastern culture on the surface needs to be adopted to the future research. Keyword : Non-Verbal Communication, Salesperson, Cross-cultural, Willing to pay(WTP), Scenario Method

Miniaturized, Flexible, Assembly-Type Wireless Communication System for Health and Environment Monitoring

Song, Min-su Kwangwoon University Graduate School 2021 국내석사

Development of the sensor technologies that diagnose health conditions or monitor the surrounding environment is of growing interest. For continuous monitoring data from the sensor, an appropriate wireless communication system is required. A lot of convention products are developed, while they have some limitations such as hard and bulky structures. Especially, whenever the type of sensor is changed, the system that process and transmit signal should be newly constructed. In this thesis, the device and assembly type concepts for miniaturized and flexible platforms that can measure the bio-signal and surrounding environment are suggested. These concepts are applied to the wireless communication system that exchanges the signal and data with sensors. They have high compatibility and a variety of options and modes to use. In this study, the wireless communication system is verified through combining with the pulse oximeter, UV, gyro, and temperature sensor that provide quantitative data via appropriate post-processing by the microcontroller. The results reported here show the possibility of application to various fields such as health care, internet of things (IoT), especially wearable devices

Performance of an adaptive modulation scheme based on deep learning

Lee, Donggu Kwangwoon University Graduate School 2021 국내석사

가변적인 채널 특성을 갖는 무선 통신 채널환경에서, 송신효율 증대를 위한 방법으로 적응형 변복조 기술은 많은 통신 시스템에 적용되고 있다. 또한 딥러닝 기법을 적응형 변복조 기술에 활용하여 기존의 수치해석 활용 방식보다 좋은 성능을 나타내는 연구들이 많이 보고되고 있다. 특히 주어진 목적에 따라 최적의 선택을 하도록 모델을 학습시키는 강화학습 기법이 적용된 연구가 있지만, 해당 연구에 따르면 연산 복잡도와 변복조 방식의 세밀한 조정 간의 딜레마와 그로 인한 성능저하가 존재한다고 한다. 따라서, 본 논문에서는 채널의 신호대비 잡음비 영역을 나누는 변복조 방식 할당 경계값을 설정하고, 이를 조절하는 심층 Q 네트워크 모델기반 적응형 변복조 알고리즘을 제안하였다. 딥러닝 모델은 제한된 비트 에러 확률 조건 내에서 스펙트럼 효율을 최대화하기 위해 경계값을 조절하여 최적의 변복조 방식 경계값을 도출한다. 본 논문에서 제안하는 경계값 조절 모델을 적용함으로써, 기존 강화학습 모델의 딜레마와 성능저하의 영향을 줄일 수 있다. 또한 본 논문의 실험 결과를 통해, 레일리 채널 환경의 광범위한 신호대비 잡음비 영역에서 평균 0.5395 bps/Hz의 스펙트럼 효율 개선 효과를 확인하였다. Adaptive modulation schemes have been actively applied in various communication systems to improve transmission rate over variable wireless communication channels. Also, there are many types of researches that outperforms conventional numerical approaches by employing deep learning-based adaptive modulation schemes. Especially, reinforcement learning, which aims to make the optimal decision, can be a strong candidate. However, according to a research using reinforcement learning algorithm, there is a dilemma between computational complexity and precise allocation of modulation schemes, and it is observed that performance degradation from the dilemma. Thus, in this paper, we propose a deep Q network-based adaptive modulation scheme that controls rate region boundaries, which divide the signal-to-noise ratio (SNR) range into symbol per bits rate region. The deep learning model derives the optimal rate region boundaries to maximize spectral efficiency within a bit error rate constraint. The effect of the dilemma and performance degradation can be mitigated with the proposed scheme. Moreover, from the simulation results, we can confirm that the proposed model improves spectral efficiency by 0.5395 bps/Hz over a wide range of SNR of the Rayleigh channel environment

Exploration of Resolution-Enhanced Three-Dimensional Images in Integral Imaging

Jang, Miao Kwangwoon University Graduate School 2012 국내박사

Three-dimensional imaging systems offer significant benefits over conventional two-dimensional imaging systems. The presence of natural depth information in the scene affords the viewer an overall improved sense of reality and naturalness. A variety of three-dimensional imaging systems attempting to reach this goal have been designed by many independent research groups, although the three-dimensional image displayed by these stereoscopic systems lack a sense of solidity and have been shown to cause viewer fatigue due to a physiological conflict resulting from supplying disparate planar images to the eyes. Integral imaging, devised by Lippmann in 1908, has been known to be able to overcome these limitations. It encodes a volume spatial model of a scene on a plane surface, and provides full color and continuous parallax both horizontally and vertically. The low resolution of integral images, however, has been always a nuisance in many applications. Hence improvement on resolution of integral images is necessary. The objective for the work in this dissertation is to develop mathematical and computational methods for resolution enhancement of three-dimensional images reconstructed under various circumstances in integral imaging. The dissertation concentrates on three representative types of circumstance: 1) resolution-enhanced magnification of far 3-D objects, 2) visibility-enhanced reconstruction of partially occluded 3-D objects, and 3) visibility-enhanced reconstruction of 3-D objects under the heavily scattering medium of dense fog. Firstly, one approach is described for resolution-enhanced magnification of far 3-D objects by employing the moving-direct-pixel-mapping method in the moving-arraylenslet-technique-based scalable integral imaging system. Optically experimental results are presented which show that greatly improved viewing-quality of the magnified 3-D reconstructed image can be obtained in the proposed system compared to the conventional moving-array-lenslet-technique-based approach. Furthermore Computational simulation results also confirm that resolution of magnified 3-D reconstructed images can be improved. Secondly, two approaches are described for visibility-enhanced reconstruction of partially occluded 3-D objects. Both approaches are developed for complex 3-D scenes in which the foreground objects may obstruct our views of the background objects. One uses the pixel-restoration scheme to minimize the effect of the occluded region on the 3-D object in each sub-image. The other one makes use of a hybrid method of smart pixel mapping and sub-image transform to computationally transform the EIA picked up from a far 3-D object into the EIA which looks like the ones visually picked up from a near 3-D object. Simulation results from both approaches reveal that the proposed approaches can improve the visual quality of the reconstructed target images, and develop the applications of integral imaging. Finally, an approach is presented for visibility-enhanced reconstruction of the 3-D object under the heavily scattering medium of dense fog. The proposed approach makes use of combination of the intermediate-view reconstruction, multi-pixel extraction and histogram equalization methods in the computational integral imaging system. By applying the HE algorithm to the resolution-enhanced SIA, a new SIA with a uniform histogram can be generated in a way that improvements of the imagecontrast of the SIA are achieved by weakening the effect of dense fog on the object image. Experimental results confirm the feasibility of the proposed method and its possibility to the practical applications including the real-time detection and recognition of the 3-D object in the heavily scattering medium.